An Algorithm for Feature Extraction of Weld Groove based on Laser Vision

Abstract

The extraction of weld groove geometry feature information is the premise of weld tracking technology. The existing seam feature information extraction technology has a good effect on thin plate, but its extraction accuracy is low when dealing with thick plate and deep weld groove. Based on the point scanning laser sensor, a method of image information preprocessing and feature extraction for thick plate is proposed, and the related algorithms are optimized. Aiming at the common right angle and oblique inflection point in the weld groove, the slope extreme value method and the oblique intercept method are used to extract the weld seam feature points respectively. Finally, the effectiveness of the proposed algorithm is verified by experiments.